It's cool how even though there is a lot of noise surrounding advice about startups a bunch of precepts have an air of unanimity. "Release early and iterate" would be one of those.
Here's a blog post by Aaron Swartz: http://www.aaronsw.com/weblog/rlrr
In there, he proposes that "release early, release often" works in open source, but is less applicable to a wide public who are not expecting something unpolished and incomplete. When you only get one chance to make an impression, the unveiling might need a lot more consideration. Of course, then your product might never even see the light of day because it's been tied up for so long.
This is a different view from that in PG's essay, and it just shows the added complexity in decision-making a startup must go through.
